System and method for managing workloads and hot-swapping a co-processor of an information handling system

• US 10,216,681 B2
• Filed: 12/01/2015
• Issued: 02/26/2019
• Est. Priority Date: 12/01/2015
• Status: Active Grant

• Current Assignee: Dell Products LP (Dell Technologies Inc.)
• Original Assignee: Dell Products LP (Dell Technologies Inc.)
• Inventors: Ahmed, Sajjad, Masuyama, Jinsaku, Palmer, John R., Ragupathi, Dinesh Kunnathur
• Primary Examiner(s): Tsai, Henry
• Assistant Examiner(s): Hassan, Aurangzeb
• Application Number:
  US14/955,977
• Publication Number:
  US 20170154005A1
• Time in Patent Office: 1,183 Days
• Field of Search: None
• CPC Class Codes: G06F 13/36 for access to common bus or bus system
• CPC Class Codes: G06F 13/4081 Live connection to bus, e.g. hot-plugging current or voltage limitation during live insertion H02H9/004
• CPC Class Codes: G06F 9/4411 Configuring for operating with peripheral devices; Loading of device drivers
• CPC Class Codes: H04L 41/0213 Standardised network management protocols, e.g. simple network management protocol [SNMP]

Abstract

An information handling system includes a host processing complex and a wireless management system. The host processing complex instantiates a hosted processing environment and includes a first general-purpose processing unit (GPU) and a GPU hot-plug module that enables a hot-plug operation to replace the first GPU with a second GPU while power is provided to the host processing complex. The hosted processing environment instantiates a first workload on the first GPU. The wireless management system operates out of band from the hosted processing environment, directs the hosted processing environment to halt the first workload, and directs the GPU hot-plug module to perform the hot-plug operation.

Claims

1. An information handling system, comprising:

a host processing complex to instantiate a hosted processing environment and including a first general-purpose processing unit (GPU) and a GPU hot-plug module that enables a hot-plug operation to replace the first GPU with a second GPU while power is provided to the host processing complex, wherein the hosted processing environment includes a plurality of workloads that can be instantiated on the first GPU, the plurality of workloads including a first workload and a second workload, and wherein the hosted processing environment instantiates the first workload on the first GPU; and

a wireless management system that operates out of band from the hosted processing environment, that directs the hosted processing environment to halt the first workload, and that directs the GPU hot-plug module to perform the hot-plug operation, and directs the hosted processing environment to launch the second workload on the second GPU after the GPU hot-plug module performs the hot-plug operation;

wherein the hosted processing environment provides a list of the workloads to the wireless management system.

2. The information handling system of claim 1, wherein the wireless management system further couples to a mobile device, provides the list to the mobile device, and receives a selection of the second workload from the mobile device, and wherein directing the hosted processing environment to launch the second workload is in response to receiving the selection.

3. The information handling system of claim 1, wherein the host processing complex includes a plurality of GPUs including the first GPU.

4. The information handling system of claim 3, wherein the hosted processing environment provides a list of the plurality of GPUs to the wireless management system.

5. The information handling system of claim 4, wherein the wireless management system further couples to a mobile device, provides the list to the mobile device, and receives a selection of the first GPU from the mobile device, and wherein directing the GPU hot-plug module to perform the hot-plug operation is in response to receiving the selection.

6. A method, comprising:

instantiating, on a host processing complex of an information handling system, a hosted processing environment;

instantiating, by the hosted processing environment, a first workload on a first general-purpose processing unit (GPU) of the host processing complex, wherein the hosted processing environment includes a plurality of workloads that can be instantiated on the first GPU, the plurality of workloads including a first workload and a second workload;

operating a wireless management system of the information handling system out of band from the hosted processing environment;

directing, by the wireless management system, the hosted processing environment to halt the first workload;

directing a GPU hot-plug module to perform a hot-plug operation to replace the first GPU with a second GPU while power is provided to the host processing complex; and

directing, by the management system, the hosted processing environment to launch a second workload on the second GPU after the GPU hot-plug module performs the hot-plug operation.

7. The method of claim 6, further comprising:

providing, by the hosted processing environment, a list of the plurality of workloads to the wireless management system.

8. The method of claim 7, further comprising:

coupling the wireless management system to a mobile device;

providing, by the wireless management system, the list to the mobile device; and

receiving, by the wireless management system a selection of the second workload from the mobile device, wherein directing the hosted processing environment to launch the second workload is in response to receiving the selection.

9. The method of claim 6, wherein the host processing complex includes a plurality of GPUs including the first GPU.

10. The method of claim 9, further comprising:

providing, by the hosted processing environment, a list of the plurality of GPUs to the wireless management system.

11. The method of claim 10, further comprising:

coupling the wireless management system to a mobile device;

providing, by the wireless management system, the list to the mobile device; and

receiving, by the wireless management system, a selection of the first GPU from the mobile device, wherein directing the GPU hot-plug module to perform the hot-plug operation is in response to receiving the selection.

12. A non-transitory computer-readable medium including code for performing a method, the method comprising:

instantiating, on a host processing complex of an information handling system, a hosted processing environment;

instantiating, by the hosted processing environment, a first workload on a first general-purpose processing unit (GPU) of the host processing complex, wherein the hosted processing environment includes a plurality of workloads that can be instantiated on the first GPU, the plurality of workloads including a first workload and a second workload;

operating a wireless management system of the information handling system out of band from the hosted processing environment;

directing, by the wireless management system, the hosted processing environment to halt the first workload;

directing a GPU hot-plug module to perform a hot-plug operation to replace the first GPU with a second GPU while power is provided to the host processing complex; and

directing the hosted processing environment to launch a second workload on the second GPU after the GPU hot-plug module performs the hot-plug operation.

13. The computer-readable medium of claim 12, further comprising:

providing, by the hosted processing environment, a list of the plurality of workloads to the wireless management system.

14. The computer-readable medium of claim 13, wherein the host processing complex includes a plurality of GPUs including the first GPU, the method further comprising:

providing, by the hosted processing environment, a list of the plurality of GPUs to the wireless management system.